Less Wrong Rationality and Mainstream Philosophy

Part of the se­quence: Ra­tion­al­ity and Philosophy

De­spite Yud­kowsky’s dis­taste for main­stream philos­o­phy, Less Wrong is largely a philos­o­phy blog. Ma­jor top­ics in­clude episte­mol­ogy, philos­o­phy of lan­guage, free will, meta­physics, metaethics, nor­ma­tive ethics, ma­chine ethics, ax­iol­ogy, philos­o­phy of mind, and more.

More­over, stan­dard Less Wrong po­si­tions on philo­soph­i­cal mat­ters have been stan­dard po­si­tions in a move­ment within main­stream philos­o­phy for half a cen­tury. That move­ment is some­times called “Quinean nat­u­ral­ism” af­ter Har­vard’s W.V. Quine, who ar­tic­u­lated the Less Wrong ap­proach to philos­o­phy in the 1960s. Quine was one of the most in­fluen­tial philoso­phers of the last 200 years, so I’m not talk­ing about an ob­scure move­ment in philos­o­phy.

Let us sur­vey the con­nec­tions. Quine thought that philos­o­phy was con­tin­u­ous with sci­ence—and where it wasn’t, it was bad philos­o­phy. He em­braced em­piri­cism and re­duc­tion­ism. He re­jected the no­tion of liber­tar­ian free will. He re­garded post­mod­ernism as sophistry. Like Wittgen­stein and Yud­kowsky, Quine didn’t try to straight­for­wardly solve tra­di­tional Big Ques­tions as much as he ei­ther dis­solved those ques­tions or re­framed them such that they could be solved. He dis­missed end­less se­man­tic ar­gu­ments about the mean­ing of vague terms like knowl­edge. He re­jected a pri­ori knowl­edge. He re­jected the no­tion of priv­ileged philo­soph­i­cal in­sight: knowl­edge comes from or­di­nary knowl­edge, as best re­fined by sci­ence. Eliezer once said that philos­o­phy should be about cog­ni­tive sci­ence, and Quine would agree. Quine fa­mously wrote:

The stim­u­la­tion of his sen­sory re­cep­tors is all the ev­i­dence any­body has had to go on, ul­ti­mately, in ar­riv­ing at his pic­ture of the world. Why not just see how this con­struc­tion re­ally pro­ceeds? Why not set­tle for psy­chol­ogy?

But isn’t this us­ing sci­ence to jus­tify sci­ence? Isn’t that cir­cu­lar? Not quite, say Quine and Yud­kowsky. It is merely “re­flect­ing on your mind’s de­gree of trust­wor­thi­ness, us­ing your cur­rent mind as op­posed to some­thing else.” Luck­ily, the brain is the lens that sees its flaws. And thus, says Quine:

Episte­mol­ogy, or some­thing like it, sim­ply falls into place as a chap­ter of psy­chol­ogy and hence of nat­u­ral sci­ence.

Yud­kowsky once wrote, “If there’s any cen­tral­ized repos­i­tory of re­duc­tion­ist-grade nat­u­ral­is­tic cog­ni­tive philos­o­phy, I’ve never heard men­tion of it.”

When I read that I thought: What? That’s Quinean nat­u­ral­ism! That’s Korn­blith and Stich and Bickle and the Church­lands and Tha­gard and Met­z­inger and Northoff! There are hun­dreds of philoso­phers who do that!

Non-Quinean philosophy

But I should also men­tion that LW philos­o­phy /​ Quinean nat­u­ral­ism is not the largest strain of main­stream philos­o­phy. Most philos­o­phy is still done in rel­a­tive ig­no­rance (or ig­nor­ing) of cog­ni­tive sci­ence. Con­sider the pref­ace to Re­think­ing In­tu­ition:

Per­haps more than any other in­tel­lec­tual dis­ci­pline, philo­soph­i­cal in­quiry is driven by in­tu­itive judg­ments, that is, by what “we would say” or by what seems true to the in­quirer. For most of philo­soph­i­cal the­o­riz­ing and de­bate, in­tu­itions serve as some­thing like a source of ev­i­dence that can be used to defend or at­tack par­tic­u­lar philo­soph­i­cal po­si­tions.

One clear ex­am­ple of this is a tra­di­tional philo­soph­i­cal en­ter­prise com­monly known as con­cep­tual anal­y­sis. Any­one fa­mil­iar with Plato’s di­alogues knows how this type of in­quiry is con­ducted. We see Socrates en­counter some­one who claims to have figured out the true essence of some ab­stract no­tion… the per­son puts for­ward a defi­ni­tion or anal­y­sis of the no­tion in the form of nec­es­sary and suffi­cient con­di­tions that are thought to cap­ture all and only in­stances of the con­cept in ques­tion. Socrates then re­futes his in­ter­locu­tor’s defi­ni­tion of the con­cept by point­ing out var­i­ous coun­terex­am­ples...

For ex­am­ple, in Book I of the Repub­lic, when Cephalus defines jus­tice in a way that re­quires the re­turn­ing of prop­erty and to­tal hon­esty, Socrates re­sponds by point­ing out that it would be un­just to re­turn weapons to a per­son who had gone mad or to tell the whole truth to such a per­son. What is the sta­tus of these claims that cer­tain be­hav­iors would be un­just in the cir­cum­stances de­scribed? Socrates does not ar­gue for them in any way. They seem to be no more than spon­ta­neous judg­ments rep­re­sent­ing “com­mon sense” or “what we would say.” So it would seem that the pro­posed anal­y­sis is re­jected be­cause it fails to cap­ture our in­tu­itive judg­ments about the na­ture of jus­tice.

After a pro­posed anal­y­sis or defi­ni­tion is over­turned by an in­tu­itive coun­terex­am­ple, the idea is to re­vise or re­place the anal­y­sis with one that is not sub­ject to the coun­terex­am­ple. Coun­terex­am­ples to the new anal­y­sis are sought, the anal­y­sis re­vised if any coun­terex­am­ples are found, and so on...

Re­fu­ta­tions by in­tu­itive coun­terex­am­ples figure as promi­nently in to­day’s philo­soph­i­cal jour­nals as they did in Plato’s di­alogues...

...philoso­phers have con­tinued to rely heav­ily upon in­tu­itive judg­ments in pretty much the way they always have. And they con­tinue to use them in the ab­sence of any well ar­tic­u­lated, gen­er­ally ac­cepted ac­count of in­tu­itive judg­ment—in par­tic­u­lar, an ac­count that es­tab­lishes their epistemic cre­den­tials.

How­ever, what ap­pear to be se­ri­ous new challenges to the way in­tu­itions are em­ployed have re­cently emerged from an un­ex­pected quar­ter—em­piri­cal re­search in cog­ni­tive psy­chol­ogy.

With re­spect to the tra­di­tion of seek­ing defi­ni­tions or con­cep­tual analy­ses that are im­mune to coun­terex­am­ple, the challenge is based on the work of psy­chol­o­gists study­ing the na­ture of con­cepts and cat­e­go­riza­tion of judg­ments. (See, e.g., Rosch 1978; Rosch and Mervis 1975; Rips 1975; Smith and Medin 1981). Psy­chol­o­gists work­ing in this area have been pushed to aban­don the view that we rep­re­sent con­cepts with sim­ple sets of nec­es­sary and suffi­cient con­di­tions. The data seem to show that, ex­cept for some math­e­mat­i­cal and ge­o­met­ri­cal con­cepts, it is not pos­si­ble to use sim­ple sets of con­di­tions to cap­ture the in­tu­itive judg­ments peo­ple make re­gard­ing what falls un­der a given con­cept...

With re­gard to the use of in­tu­itive judg­ments ex­em­plified by re­flec­tive equil­ibrium, the challenge from cog­ni­tive psy­chol­ogy stems pri­mar­ily from stud­ies of in­fer­ence strate­gies and be­lief re­vi­sion. (See, e.g., Nis­bett and Ross 1980; Kah­ne­man, Slovic, and Tver­sky 1982.) Numer­ous stud­ies of the pat­terns of in­duc­tive in­fer­ence peo­ple use and judge to be in­tu­itively plau­si­ble have re­vealed that peo­ple are prone to com­mit var­i­ous fal­la­cies. More­over, they con­tinue to find these fal­la­cious pat­terns of rea­son­ing to be in­tu­itively ac­cept­able upon re­flec­tion… Similarly, stud­ies of the “in­tu­itive” heuris­tics or­di­nary peo­ple ac­cept re­veal var­i­ous gross de­par­tures from em­piri­cally cor­rect prin­ci­ples...

There is a grow­ing con­sen­sus among philoso­phers that there is a se­ri­ous and fun­da­men­tal prob­lem here that needs to be ad­dressed. In fact, we do not think it is an over­state­ment to say that Western an­a­lytic philos­o­phy is, in many re­spects, un­der­go­ing a crisis where there is con­sid­er­able ur­gency and anx­iety re­gard­ing the sta­tus of in­tu­itive anal­y­sis.


So Less Wrong-style philos­o­phy is part of a move­ment within main­stream philos­o­phy to mas­sively re­form philos­o­phy in light of re­cent cog­ni­tive sci­ence—a move­ment that has been ac­tive for at least two decades. More­over, Less Wrong-style philos­o­phy has its roots in Quinean nat­u­ral­ism from fifty years ago.

And I haven’t even cov­ered all the work in for­mal episte­mol­ogy to­ward (1) math­e­mat­i­cally for­mal­iz­ing con­cepts re­lated to in­duc­tion, be­lief, choice, and ac­tion, and (2) ar­gu­ing about the foun­da­tions of prob­a­bil­ity, statis­tics, game the­ory, de­ci­sion the­ory, and al­gorith­mic learn­ing the­ory.

So: Ra­tion­al­ists need not dis­miss or avoid philos­o­phy.

Up­date: To be clear, though, I don’t recom­mend read­ing Quine. Most peo­ple should not spend their time read­ing even Quinean philos­o­phy; learn­ing statis­tics and AI and cog­ni­tive sci­ence will be far more use­ful. All I’m say­ing is that main­stream philos­o­phy, es­pe­cially Quinean philos­o­phy, does make some use­ful con­tri­bu­tions. I’ve listed more than 20 of main­stream philos­o­phy’s use­ful con­tri­bu­tions here, in­clud­ing sev­eral in­stances of clas­sic LW dis­solu­tion-to-al­gorithm.

But maybe it’s a tes­ta­ment to the epistemic util­ity of Less Wrong-ian ra­tio­nal­ity train­ing and think­ing like an AI re­searcher that Less Wrong got so many things right with­out much in­ter­ac­tion with Quinean nat­u­ral­ism. As Daniel Den­nett (2006) said, “AI makes philos­o­phy hon­est.”

Next post: Philos­o­phy: A Diseased Discipline


Den­nett (2006). Com­put­ers as Pros­the­ses for the Imag­i­na­tion. Talk pre­sented at the In­ter­na­tional Com­put­ers and Philos­o­phy Con­fer­ence, Laval, France, May 3, 2006.

Kah­ne­man, Slovic, & Tver­sky (1982). Judg­ment Un­der Uncer­tainty: Heuris­tics and Bi­ases. Cam­bridge Univer­sity Press.

Nis­bett and Ross (1980). Hu­man In­fer­ence: Strate­gies and Short­com­ings of So­cial Judg­ment. Pren­tice-Hall.

Rips (1975). In­duc­tive judg­ments about nat­u­ral cat­e­gories. Jour­nal of Ver­bal Learn­ing and Be­hav­ior, 12: 1-20.

Rosch (1978). Prin­ci­ples of cat­e­go­riza­tion. In Rosch & Lloyd (eds.), Cog­ni­tion and Cat­e­go­riza­tion (pp. 27-48). Lawrence Er­lbaum As­so­ci­ates.

Rosch & Mervis (1975). Fam­ily re­sem­blances: stud­ies in the in­ter­nal struc­ture of cat­e­gories. Cog­ni­tive Psy­chol­ogy, 8: 382-439.

Smith & Medin (1981). Con­cepts and Cat­e­gories. MIT Press.